Can anyone suggest a switch for me ?

[Dave]

Hi,
I will have 4 inputs, all will be 10 kHz sine waves (perhaps very noisy)
and I want to connect a measuring instrument to each in turn.

This is to go into a vector network analyser I am designing. I have not
sat down carefully and worked out what the limiting item on performance
will be, but I suspect I will need 60 dB isolation to prevent this
section limiting performance. So if I am measuring input 2, the
contribution from inputs 1, 2 and 4 should be at least 60 dB down on
that from port 2.

 

[Dave]

Dave wrote:

Hi,
I will have 4 inputs, all will be 10 kHz sine waves (perhaps very noisy)
and I want to connect a measuring instrument to each in turn.

This is to go into a vector network analyser I am designing. I have not
sat down carefully and worked out what the limiting item on performance
will be, but I suspect I will need 60 dB isolation to prevent this
section limiting performance. So if I am measuring input 2, the
contribution from inputs 1, 2 and 4 should be at least 60 dB down on
that from port 2.

I should have added I would rather soemthing that can be switched

electronically (not relays), but perhaps relays are the only way.

Maximum signal will be 1V peak. minimum might be a few uV.

 

[Ken Smith]

In article <42f808ba@212.67.96.135>, Dave <nospam@nowhere.com> wrote:

Dave wrote:
Hi,
I will have 4 inputs, all will be 10 kHz sine waves (perhaps very noisy)
and I want to connect a measuring instrument to each in turn.

This is to go into a vector network analyser I am designing. I have not
sat down carefully and worked out what the limiting item on performance
will be, but I suspect I will need 60 dB isolation to prevent this
section limiting performance. So if I am measuring input 2, the
contribution from inputs 1, 2 and 4 should be at least 60 dB down on
that from port 2.

I should have added I would rather soemthing that can be switched
electronically (not relays), but perhaps relays are the only way.

Maximum signal will be 1V peak. minimum might be a few uV.

Since you didn't say anything about needing to switch quickly, I'll assume
slow is ok. If you use CD4016 chips arranged like the below, the cross
talk will be determined more by the layout than the devices.


U1A U2A
CH1 ------o\----+-------o\-------+------ To output
! !
o !
/ U1B !
! !
GND !
U3A U2B !
CH2 ------o\----+-------o\-------+
! !
o !
/ U3B !
! !
GND !
more channels
--
--
kensmith@rahul.net forging knowledge

 

[Dave]

Ken Smith wrote:

Since you didn't say anything about needing to switch quickly, I'll assume
slow is ok. If you use CD4016 chips arranged like the below, the cross
talk will be determined more by the layout than the devices.


U1A U2A
CH1 ------o\----+-------o\-------+------ To output
! !
o !
/ U1B !
! !
GND !
U3A U2B !
CH2 ------o\----+-------o\-------+
! !
o !
/ U3B !
! !
GND !
more channels

As your guessed, speed it not critical. I'll take a look at thw data on
that chip.